architecture.

The rapid advancement of digital technologies has fundamentally transformed conventional urban environments into intelligent and interconnected smart cities. The convergence of the Internet of Things, cloud computing, artificial intelligence, and advanced communication technologies has enabled the deployment of a wide range of smart urban services, including intelligent transportation systems, smart healthcare, environmental monitoring, public safety, and energy management. These services are designed to improve urban sustainability, optimize resource utilization, and enhance citizens’ quality of life through continuous sensing, data-driven decision-making, and intelligent service automation [1-3].

The large-scale deployment of IoT devices has become a fundamental component of modern smart city infrastructures. Millions of heterogeneous devices, including environmental sensors, surveillance cameras, wearable devices, connected vehicles, and smart meters, continuously generate massive volumes of data characterized by high velocity, variety, and temporal dynamics. Efficient processing of these data streams is essential for supporting latency-sensitive applications that require immediate responses, such as intelligent traffic control, emergency management, industrial automation, and remote healthcare monitoring. Consequently, smart city infrastructures require distributed computing platforms capable of delivering scalable, reliable, and real-time data processing while maintaining efficient utilization of communication and computational resources [4-6].

Cloud computing has long been considered the primary computational platform for IoT applications due to its virtually unlimited processing capability, elastic resource provisioning, and centralized storage infrastructure. However, transmitting data from geographically distributed IoT devices to remote cloud data centers introduces considerable communication latency, increases bandwidth consumption, and may create network congestion and privacy concerns. These limitations significantly reduce the effectiveness of cloud-centric architectures for applications requiring real-time decision-making and ultra-low response times [7-9]. Fog computing has emerged as an effective distributed computing paradigm that extends cloud capabilities toward the network edge by deploying intermediate computational resources between IoT devices and centralized cloud infrastructures. Instead of transmitting all generated data to remote cloud servers, fog nodes perform localized data filtering, preprocessing, aggregation, temporary storage, and real-time analytics near data sources. This hierarchical processing model substantially reduces communication latency, decreases unnecessary network traffic, improves Quality of Service, and enhances system responsiveness. Moreover, fog computing provides a collaborative computing environment in which IoT devices, fog nodes, and cloud platforms cooperate to efficiently execute computational workloads according to application requirements [1], [2], [7].

During recent years, research on fog computing has evolved from conceptual architectures toward intelligent resource management, application placement, task scheduling, service orchestration, and AI-assisted optimization techniques. Existing studies have demonstrated that fog computing can significantly improve latency, bandwidth efficiency, and resource utilization in distributed IoT environments. Nevertheless, most available approaches focus on specific optimization problems or individual application scenarios, whereas relatively few studies investigate integrated hierarchical architectures capable of simultaneously coordinating IoT devices, fog nodes, and cloud infrastructures in large-scale smart city environments [8].

Despite the considerable progress achieved in recent years, several research challenges remain unresolved. Efficient workload distribution, adaptive resource allocation, interoperability among heterogeneous IoT devices, cybersecurity, and intelligent orchestration continue to limit the large-scale deployment of fog computing in smart cities. Furthermore, balancing computational workloads between fog and cloud layers while satisfying the stringent latency, scalability, and reliability requirements of next-generation urban applications remains an open research challenge [9], [10].

Motivated by these challenges, this study proposes a hierarchical fog computing framework for low-latency data processing in smart city IoT systems. The proposed framework consists of three collaborative computational layers: the IoT layer, responsible for real-time data acquisition; the fog layer, responsible for localized processing and latency-sensitive decision-making; and the cloud layer, responsible for large-scale analytics, historical data management, and computationally intensive processing. By intelligently distributing computational tasks across these layers, the proposed framework aims to reduce end-to-end latency, improve bandwidth utilization, enhance resource efficiency, and increase the scalability of smart city services.

The main contributions of this study are summarized as follows:

I. A hierarchical IoT–Fog–Cloud architecture is proposed to support scalable and low-latency data processing in heterogeneous smart city environments.

II. A distributed workload allocation strategy is introduced to execute latency-sensitive tasks at fog nodes while assigning computation-intensive operations to cloud infrastructures.

III. The proposed framework is analyzed according to key system performance indicators, including latency, bandwidth utilization, scalability, resource efficiency, and service reliability.

IV. Current research challenges and future research directions are discussed, focusing on artificial intelligence-assisted resource management, cybersecurity, interoperability, and the integration of fog computing with emerging communication technologies.

The remainder of this paper is organized as follows. Section 2 reviews recent advances in fog computing and smart city IoT systems. Section 3 presents the proposed hierarchical architecture and its operational workflow. Section 4 describes the research methodology adopted in this study. Section 5 discusses the analytical evaluation of the proposed framework. Section 6 highlights current challenges and future research opportunities, while Section 7 concludes the paper.

Fog computing has become one of the most promising distributed computing paradigms for addressing the limitations of cloud-centric architectures in Internet of Things environments. By extending computational and storage resources from centralized cloud data centers to the network edge, fog computing enables data processing closer to IoT devices, thereby reducing communication latency and improving the responsiveness of time-critical applications. The concept was initially introduced by Bonomi et al. [1], who demonstrated that deploying computational resources near data sources could significantly enhance the performance of latency-sensitive IoT services. Subsequently, Dastjerdi and Buyya [11] established fog computing as an intermediate computing layer between IoT devices and cloud infrastructures, highlighting its capability to support scalable and geographically distributed IoT applications.

As fog computing matured, research gradually shifted from conceptual architectures toward comprehensive architectural frameworks and system-level management. Mahmud et al. [5] presented a detailed taxonomy of fog application management by categorizing existing approaches according to resource management, service provisioning, scalability, and application deployment. Similarly, Singh et al. [3] reviewed the evolution of fog computing architectures and identified resource allocation, interoperability, and Quality of Service provisioning as the primary technical challenges for large-scale distributed environments. Mann [12] further analyzed architectural models for fog systems and emphasized that hierarchical deployment strategies improve flexibility and coordination among distributed computational resources.

Another major research direction focuses on resource management and workload orchestration within fog environments. As the number of IoT devices continues to increase, efficient allocation of computational resources has become essential for maintaining system performance. provided a comprehensive review of resource allocation and task scheduling techniques in fog and Internet of Everything (IoE) environments, demonstrating that intelligent scheduling strategies substantially reduce service latency and improve resource utilization. More recently, Alsadie [13] investigated the integration of artificial intelligence techniques into fog resource management, showing that machine learning algorithms can improve adaptive workload distribution under dynamic network conditions. Similarly, highlighted the growing importance of intelligent resource management and identified distributed orchestration as one of the major future research directions.

Despite the considerable progress achieved during the last decade, several research gaps remain. Most existing studies primarily investigate individual optimization problems, including task scheduling, application placement, energy efficiency, or security, without providing an integrated framework capable of coordinating IoT devices, fog nodes, and cloud infrastructures simultaneously. Furthermore, many proposed solutions are evaluated within application-specific scenarios, limiting their applicability to heterogeneous smart city environments characterized by dynamic workloads, diverse communication technologies, and continuously evolving service requirements.

To address these limitations, this study proposes a hierarchical IoT-Fog-Cloud computing framework that emphasizes collaborative workload distribution across multiple computational layers. Unlike previous studies that mainly concentrate on isolated optimization strategies, the proposed framework provides an integrated architecture that combines localized fog processing with cloud-based large-scale analytics. This hierarchical design aims to reduce end-to-end latency, optimize bandwidth utilization, improve scalability, and support real-time decision-making for next-generation smart city IoT systems.

A comparative summary of recent studies on fog computing for smart city IoT applications is presented in Table 1, highlighting their major contributions and remaining research gaps

The continuous expansion of Internet of Things infrastructures in smart cities has significantly increased the demand for distributed computing platforms capable of processing heterogeneous data streams with low latency and high reliability. Modern smart city applications including intelligent transportation systems, public safety, environmental monitoring, healthcare, and smart energy management require rapid data analysis and real-time decision-making, which cannot always be efficiently supported by conventional cloud-centric architectures. The transmission of massive volumes of IoT-generated data to geographically distant cloud data centers introduces communication delays, increases bandwidth consumption, and may negatively affect the performance of latency-sensitive services.

Fog computing has emerged as an effective solution to these challenges by extending computational capabilities from centralized cloud infrastructures toward the network edge. Instead of relying exclusively on remote cloud servers, fog computing distributes computation across multiple intermediate nodes located closer to IoT devices. This distributed processing paradigm enables localized data analysis, minimizes unnecessary data transmission, and improves the responsiveness of time-critical applications while preserving the computational advantages of cloud platforms.

Based on these principles, this study proposes a hierarchical IoT - Fog-Cloud architecture designed to support efficient and low latency data processing in smart city environments. The proposed framework organizes computational resources into three collaborative layers, where each layer performs specific functions according to its computational capacity, latency requirements, and operational responsibilities. The hierarchical organization facilitates efficient workload distribution, reduces communication overhead, and enables scalable processing of heterogeneous IoT workloads.

This study adopts a conceptual and analytical research methodology to investigate the potential of fog computing for improving data processing efficiency in smart city Internet of Things environments. Rather than implementing a physical prototype or conducting simulation-based experiments, the research focuses on developing a hierarchical IoT-Fog-Cloud framework and analyzing its operational characteristics based on established principles of distributed computing and findings reported in the existing literature. The methodological approach integrates architectural design with qualitative performance analysis to examine how fog computing can address the limitations of conventional cloud-centric IoT systems.

The proposed research methodology consists of four sequential phases, as illustrated in the conceptual development process of the framework

The proposed hierarchical IoT–Fog–Cloud framework is conceptually evaluated by examining its potential to address the major limitations of conventional cloud-centric computing architectures in smart city Internet of Things (IoT) environments. Rather than relying on centralized processing, the proposed architecture distributes computational tasks across IoT devices, fog nodes, and cloud infrastructures according to application requirements and latency constraints. This distributed computing model is expected to improve processing efficiency while supporting the scalability and responsiveness required by modern smart city services.

Despite the significant advantages offered by fog computing in smart city IoT environments, several technical and operational challenges continue to limit its large-scale deployment. One of the primary concerns is resource management across distributed fog nodes. Unlike centralized cloud infrastructures, fog environments consist of heterogeneous devices with varying computational capabilities, storage capacities, and energy resources. Efficient allocation of computational tasks while maintaining balanced workloads remains a complex challenge, particularly in highly dynamic urban environments.

Another important issue is interoperability among heterogeneous IoT devices and communication protocols. Smart city infrastructures integrate a wide range of sensors, intelligent devices, and communication technologies developed by different manufacturers. Ensuring seamless communication and standardized data exchange across these heterogeneous platforms requires the adoption of unified architectural frameworks and interoperable communication standards.

Security and privacy also represent critical challenges for fog computing. Since data processing is distributed across multiple fog nodes located near end users, the attack surface becomes significantly larger than in conventional cloud environments. Protecting sensitive information against unauthorized access, cyberattacks, and data manipulation requires robust authentication mechanisms, encryption techniques, secure communication protocols, and continuous monitoring of distributed computing resources.

Scalability remains another research challenge as smart cities continue to expand. The rapid increase in connected IoT devices generates continuously growing volumes of data, requiring adaptive resource management strategies capable of maintaining consistent system performance under fluctuating workloads. Future fog computing architectures should incorporate intelligent workload scheduling and autonomous resource allocation mechanisms to improve computational efficiency while minimizing operational costs. Emerging technologies are expected to further enhance the capabilities of fog computing. Artificial intelligence and machine learning can enable intelligent resource allocation, predictive maintenance, anomaly detection, and adaptive decision-making within distributed fog environments. Similarly, the integration of fifth-generation (5G) and sixth-generation (6G) communication networks is expected to provide ultra-low latency, higher bandwidth, and more reliable connectivity, enabling more efficient collaboration between IoT devices, fog nodes, and cloud infrastructures.

Future research should focus on developing intelligent, secure, and energy efficient fog computing frameworks capable of supporting large scale smart city ecosystems. Particular attention should be given to AI-driven resource management, cybersecurity, interoperability, sustainable energy consumption, and seamless integration with next generation communication technologies. Addressing these challenges will play a fundamental role in realizing resilient, scalable, and intelligent urban infrastructures.

The rapid growth of Internet of Things technologies has fundamentally transformed the computational requirements of modern smart cities, creating an increasing demand for distributed computing architectures capable of supporting real-time data processing, low communication latency, and scalable service delivery. Although cloud computing continues to provide substantial computational power and storage capacity, its centralized architecture presents significant challenges for latency-sensitive applications due to communication delays, bandwidth limitations, and network congestion. These limitations highlight the need for complementary computing paradigms that can process data closer to where it is generated.

In this study, a hierarchical IoT-Fog-Cloud computing framework was proposed as a conceptual architecture for supporting efficient data processing in smart city environments. The framework organizes computational resources into three collaborative layers that distribute processing tasks according to latency requirements and computational complexity. By enabling localized processing at the fog layer while preserving the large-scale analytical capabilities of cloud infrastructures, the proposed architecture is designed to improve communication efficiency, reduce unnecessary data transmission, and support scalable processing of heterogeneous IoT workloads.

The conceptual analysis presented in this paper suggests that hierarchical fog computing provides an effective architectural foundation for next generation smart city systems. The proposed framework has the potential to improve bandwidth utilization, optimize distributed resource management, enhance service continuity, and support latency-sensitive applications such as intelligent transportation, environmental monitoring, public safety, healthcare systems, and smart energy management. Furthermore, the hierarchical collaboration among IoT devices, fog nodes, and cloud infrastructures offers greater flexibility for managing the continuously growing number of connected devices within large-scale urban environments.

Despite these promising characteristics, the present study is limited to the conceptual design and qualitative analysis of the proposed architecture. Consequently, quantitative performance evaluation through simulation platforms and real-world deployments remains an important direction for future research. Experimental validation using tools such as iFogSim2 or EdgeCloudSim, together with realistic smart city datasets, would enable comprehensive assessment of latency, bandwidth utilization, energy efficiency, scalability, and resource management under different operating conditions.

Future research should also investigate the integration of fog computing with emerging technologies, including artificial intelligence-driven resource orchestration, edge intelligence, digital twin platforms, software-defined networking , network function virtualization, and next-generation 5G/6G communication networks. The convergence of these technologies is expected to enable more autonomous, adaptive, secure, and sustainable smart city infrastructures capable of meeting the increasing computational demands of future urban ecosystems.

Overall, the proposed hierarchical IoT–Fog–Cloud framework provides a comprehensive conceptual foundation for distributed data processing in smart city IoT systems. By emphasizing collaborative computation, localized intelligence, and scalable workload distribution, this study contributes to the ongoing development of efficient fog computing architectures and establishes a basis for future experimental research and practical deployment in next generation intelligent urban environments.

We sincerely thank KIIT UNIVERSITY for providing the necessary resources and infrastructure to conduct this research. Special thanks go to Dr. Hitesh Mahapatra for their invaluable guidance and support throughout this project. We also thank our colleagues for their insightful feedback and technical assistance.

This research was supported by KIIT University, which provided the infrastructure and computational resources necessary for conducting the experiments and simulations. We extend our appreciation to the administrative and technical staff for their in-kind support, including access to laboratory equipment, computing facilities, and materials vital for completing this research on fog computing and smart city IoT systems.

Data Availability

No datasets were generated or analyzed during the current study. Therefore, data sharing is not applicable to this article.

Conflicts of Interest

The authors declare no conflict of interest. Funders, including KIIT University, played no role in the design of the study, the collection, analysis, or interpretation of the data, the writing of the manuscript, or the decision to publish the results. No personal circumstances or interests could be perceived as having an inappropriate influence on the presentation or interpretation of the research findings.